GlossaryAI OperationsAI Cost Ratio
AI Operations

AI Cost Ratio

AI Cost Ratio is the percentage of a company's revenue spent on AI compute — LLM API calls, inference costs, model training, and AI infrastructure. It's the new version of "cost of goods sold" for AI-run companies.

Formula

AI Cost Ratio = (Total AI compute costs / Revenue) × 100

Benchmarks

AI Cost RatioAssessment
< 5%Excellent — highly efficient AI usage
5–10%Good — standard for AI-run SaaS
10–20%Acceptable — may need optimization
20–30%Concerning — AI costs eating margins
30%+Unsustainable — rethink architecture

Typical AI Costs (2026 Pricing)

OperationCostExample
Claude Sonnet API call~$0.003–0.015 per callCustomer support response
GPT-4 API call~$0.005–0.03 per callContent generation
Image generation~$0.02–0.05 per imageMarketing visuals
Code generation~$0.01–0.05 per taskBug fix, feature
Embedding / search~$0.0001 per queryKnowledge retrieval

AI Cost Optimization

  1. Model selection — Use smaller models for simple tasks, large models for complex ones
  2. Caching — Cache frequent queries to avoid redundant API calls
  3. Batching — Process multiple requests together for volume discounts
  4. Prompt optimization — Shorter, more efficient prompts reduce token costs
  5. Self-hosted models — For high-volume tasks, running open-source models can be cheaper

AI Cost Ratio vs Traditional SaaS Costs

Cost CategoryTraditional SaaSAI-Run SaaS
Engineering25–35% of revenue2–5% (AI compute)
Marketing20–30% of revenue1–3% (AI compute)
Support10–15% of revenue0.5–2% (AI compute)
Infrastructure5–10% of revenue5–10% (same)
Total60–90%8–20%

The AI cost ratio replaces three separate human cost categories with one much smaller compute cost.

On EvolC, we track AI cost ratio as a key efficiency metric — it shows how well a company uses AI and how much margin flows to investors.

Compare AI efficiency across companies →